The Data

Here, disproportionality index is used as a measure of equitable representation. It is calculated as below.
\[\frac{\%_{Group\: X; \:salary\: range \: i}}{\% _{All \:Employees; \:salary\: range \: i}} = DI _{Group\:X; \:salary\:range\:i}\]

For example, 23.51% of Black employees and 16.70% of all employees are are within the <60K salary range. This gives us a DI of 1.41.

\[ DI_{Black \: employees; <60K } =\frac{23.51\%}{16.70\%} = 1.41\]

Below are the values used to create the figures. This data is based on self identification of equity-deserving groups, and represents pre-tax salary ranges excluding bonuses and promotions, publically availble here. It was collected in March of 2021. Please note that this data was not imputed for suppressed numbers. For more information and the scripts used to produce this analysis, visit the project’s GitHub repository.

Plot 1: DI of all groups by salary range

The dotted line represents equitable representation. It can been seen that Black employees are overrepresented at the lower salary ranges and underrepresented at higher ranges compared to the other groups. As can be seen by the “All Non-Minority Employees” bars, an equal DI across all salary ranges should be expected.

English

Français

Plot 2: DI Share by subgroups and salary range

Here, DI is subtracted from 1 to calculate a DI Share metric, meaning that 0 rather than 1 would represent equitable representation. Each column is coloured based on if the DI Share is negative or positive.

English

Français

Plot 3: A direct comparison of DI in Black Employees vs. Non-Visible Minority Employees

This plot zooms in on our “baseline” group and compares it to the Black employee group, which appears to be the most affected by inequity. A steady decrease in representation with each increase in salary range is clearly visible for the Black employee group.

English

Français